1. Auditory function and expression analysis of a partial genomic deletion allele of the mouse Coch gene. A National Cancer Institute (NCI) group led by Dr. Colin Stewart has created a Coch partial deletion mouse line in which the entire protein downstream of the LCCL domain, including two VWF domains, has been deleted by homologous recombination. We previously reported evidence that is consistent with published hypotheses that DFNA9 mutations act via a gain-of-function or dominant negative effect upon other genes, and raise the possibility that Coch itself is not required for normal hearing. Since last year's report, we have used a related lacZ reporter expression allele, and performed western blot studies, to analyze the expression of Coch in the mouse auditory system. The results indicate a role for Coch in the extracellular matrices of the nonsensory portions of the cochlea, and confirm that the partial genomic deletion allele results in loss of all detectable cochlin protein in the inner ears of homozygous mutant mice. 2. We are continuing our studies to determine the structure, function, and expression of the transmembrane channel-like (TMC) genes. We are currently generating mice segregating knockout (null) alleles of Tmc1, Tmc2, and Tmc3. We hav generated and characterized antibodies to Tmc1 and Tmc2; we have used these antibodies to determine the subcellular locations of these two proteins in inner ear sensory hair cells. We have used heterologous mammalian tissue culture systems to study the function, structure, and expression of Tmc1, Tmc2, and Tmc3. These studies have included analyses of physical interactions, documentation of specificities of antibodies, and patch-clamp studies to detect transmembrane ion movements associated with expression of these proteins. These studies indicate that Tmc1 and Tmc2 are not trafficked to the plasma membrane in these systems; this has prevented us from conventional patch-clamp studies. We are also using these systems to delineate the transmembrane topologic organization of these proteins. We have used a yeast two-hybrid screen to isolate genes encoding proteins that interact with Tmc1 and Tmc2. We are using a combination of approaches to determine which interactions occur in situ in hair cells, and have narrowed the list to a few candidate genes. We have also determined that background mouse strain has an effect on the rate and extent of outer hair cell degeneration in heterozygous Beethoven mice. We are currently genotyping the N2 offspring of a mapping backcross to identify the genetic map location(s) of the loci responsible for these differences. 3. We are continuing our efforts to positionally clone the gene mutated in the mouse Twirler strain. Heterozygous Twirler mice have inner ear malformations and obesity, whereas homozygous mice are born with cleft palate and die at birth. The critical interval containing the Twirler gene is approximately 1 Mb and contains several known genes along with several predicted genes. We have completed nucleotide sequence analysis all of the exons in these genes and not found a likely pathogenic mutation. There is an excellent candidate gene in the interval, for which there is a knockout mouse. We are now performing complementation testing with that knockout to determine if Twirler is allelic with the knockout; non-complementation would indicate that Twirler is indeed an allele of that gene. We have also obtained a transgene insertional mutant mouse, Tg9257, with a very similar phenotype to Twirler. Although the only detectable transgene insertion is several megabases away from the Twirler interval, there may be a mutation in the interval that is simply linked to the transgene. Therefore we are repeating the mutation search on Tg9257 with the genes in the Twirler interval, in the even that Tg9257 may be allelic with Twirler. 4. We have identified the first known genetic modifier of human hereditary hearing loss. Although we identified through a candidate gene approach in a single large family segregating DFNB12 hearing loss, we have preliminary evidence that it may modify hearing loss due to other genetic causes, too. We have confirmed that the putative modifier allele has a significant alteration of its wild type functional activity, and 3-D modelling confirms that the modifier missense substitution has an effect on the structure of the encoded protein. 5. Enlargement of the vestibular aqueduct (EVA) is thought to be the most commonly detected radiologic malformation in temporal bones of individuals with hearing loss. A significant proportion of EVA cases have been reported to be associated with mutations of the SLC26A4 gene, in which mutations cause Pendred syndrome. PS is an autosomal recessive disorder comprised of bilateral sensorineural hearing loss and thyroid goiter. Goiter is an incompletely penetrant feature, but the perchlorate discharge test is a comparatively sensitive and specific method to detect the underlying iodide organification defect in PS thyroid glands. EVA is a universal finding in the ears of affected PS individuals. We have now ascertained nearly 80 EVA probands at the NIH Clinical Center. We have determined that PS is correlated with two mutant SLC26A4 alleles, and nonsyndromic EVA is associated with one or zero mutant SLC26A4 alleles. Furthermore, the range of auditory phenotypes differs between the nonsyndromic EVA and PS groups. The observed correlation of thyroid and auditory phenotypes with SLC26A4 genotype suggests that most, if not all, nonsyndromic EVA is not associated with bi-allelic SLC26A4 mutations. Based upon our data, we hypothesize that one or more other genetic or environmental factors may act alone or in combination with a single SLC26A4 mutation to cause EVA. We have begun analyzing a number of candidate genes based upon known association with EVA in other hearing loss syndromes, or based upon molecular function, as candidates for the etiologic cofactor. We have collaborated with Dr. Hong-Joon Park of the Soree Ear Clinic in Seoul, Korea, to study the contribution of SLC26A4 mutations to EVA in Korean patients. We have determined that nearly all Korean patients have two detectable mutations (and thus likely have Pendred syndrome); this is a much higher proportion than in other studied populations, and may indicate that Koreans lack the other etiologic cofactors that contribute to nonsyndromic EVA in other populations.